Jingshan Huang, Rosa Laura Zavala Gutiérrez, Benito Mendoza García, and Michael N. Huhns
As the Semantic Web gains attention as the next generation of the Web, the issue of reconciling different views of independently developed and exposed data sources becomes increasingly important. Ontology integration serves as a basis for solving this problem. In this paper, we describe an approach to construct a consensus ontology from numerous, independently designed ontologies. Our method has the following features: i) the matching is carried out at the schema level; ii) the alignment of the ontologies is performed without previous agreement on the semantics of the terminology used by each ontology; iii) both the linguistic and the contextual features of an ontology concept are considered; iv) WordNet is incorporated into the linguistic analysis phase; v) heuristic knowledge is integrated into the contextual analysis phase; and vi) reasoning rules based on the domain-independent relationships subclass, superclass, equivalentclass, sibling, and each ontology concept’s property list are used to infer new relationships among concepts. We describe a set of experiments and provide an evaluation of the results that shows the accuracy of our system.